CN106597122A - Radar and communication signal pulse width detection algorithm - Google Patents

Abstract

The invention provides a pulse width detection algorithm for radar and communication signals, which first detects the position of the rising edge, then detects the position of the falling edge, and then obtains the pulse width through the position of the rising edge and the position of the falling edge, and then enters the detection of the rising edge again , so that the detection of the rising edge and the detection of the falling edge are performed cyclically, and the width of the pulse is continuously captured and calculated. When detecting the rising edge and falling edge, the combination of wide window and narrow window is adopted to balance the relationship between the detection probability and the positioning of rising and falling edges, ensuring extremely low pulse loss probability and extremely low false alarm probability. At the same time, the high-resolution capability of the narrow window is used to improve the detection accuracy of the rising edge and the falling edge; on the other hand, the present invention adopts a two-level decision fusion mechanism, and the fusion result after multiple narrow window decision fusion is combined with the wide window The fusion of judgment results further improves the detection probability and detection accuracy.

Description

A Pulse Width Detection Algorithm for Radar and Communication Signals

technical field

The invention relates to a pulse width detection algorithm for radar and communication signals, and belongs to the technical fields of radar pulse signal detection and communication pulse signal detection.

Background technique

Under the conditions of modern warfare, electronic countermeasures and the struggle against countermeasures are becoming increasingly fierce, so effectively intercepting, sorting and identifying radar and communication signals are the main tasks of passive receiving weapon systems. In the existing electromagnetic environment, a large number of various pulse signals appear every second. As one of the five core parameters of electronic reconnaissance, pulse width plays a vital role in radar and communication signal analysis. Therefore, the accurate measurement of the pulse width of the radar signal is one of the tasks to be solved urgently in modern signal processing.

Because pulse width detection needs to be completed in a very short time, many advanced signal processing techniques are difficult to adopt. The energy detection method is the most widely used pulse width detection method in the prior art. This method calculates the average power or energy of the sampling point within a period of time, and then judges the rising edge and falling edge according to the relationship between the average power or energy and the threshold value. Whether the edge appears, the computational complexity is low, and there is no need to know the prior information of the signal, so it is widely used. However, in the actual application process, if a wide window is used, that is, there are more sampling points in the window, the detection probability of the energy detection method is high under the same false alarm probability, but it is difficult to accurately locate the rising and falling edges; if a narrow window is used That is, if the number of sampling points in the window is small, it is convenient to accurately locate the rising edge and falling edge under the same false alarm probability, but the detection probability is very low, and errors are prone to occur. Therefore, the root mean square error of the detection result of the energy detection method is relatively large, and the measurement accuracy of the pulse width is low.

Contents of the invention

The present invention proposes a pulse width detection algorithm for radar and communication signals, the purpose of which is to: (1) Improve the precision of pulse width detection on the basis of low computational complexity.

Technical scheme of the present invention is as follows:

A pulse width detection algorithm for radar and communication signals, the steps are:

First detect the position of the rising edge, then detect the position of the falling edge, then obtain the pulse width through the position of the rising edge and the position of the falling edge, and then enter the detection of the rising edge again, so as to cycle the detection of the rising edge and the detection of the falling edge, and continuously capture Pulse signal and calculate pulse width, characterized by:

Before detection, first establish a wide time window, then divide the wide time window into several narrow time windows, and set a wide time window threshold for the wide time window, and set a high threshold and a low threshold for the narrow time window;

When detecting the rising edge, firstly judge the over-high threshold and the low-threshold judgment of the narrow time window, then carry out the high-threshold decision fusion and the low-threshold decision fusion, and then carry out the wide time window cross-threshold judgment and get the "wide time window rising edge position", and then perform second-level decision fusion on the wide time window detection result and the rising edge low threshold fusion decision to obtain the "low threshold rising edge position", and perform the second level decision on the wide time window detection result and the rising edge high threshold fusion decision Decision fusion and obtain the "high threshold rising edge position", and estimate the signal-to-noise ratio of the sampling points that pass the threshold in a wide time window. Finally, according to the result of the signal-to-noise ratio estimation, the " and "Wide Time Window Rising Edge Position" select one as "Rising Edge Position";

When detecting the falling edge, first judge the over-high threshold and the low-threshold judgment of the narrow time window, then perform the high-threshold decision fusion and low-threshold decision fusion, and then carry out the wide time window cross-threshold judgment to complete the wide time window power greater than or equal to The signal-to-noise ratio of the sampling point of the threshold is estimated to obtain the "wide time window falling edge position", and then the wide time window detection result is combined with the falling edge low threshold fusion decision for the second-level decision fusion to obtain the "low threshold falling edge position", The second-level decision fusion is performed on the wide time window detection result and the falling edge high threshold fusion decision to obtain the "high threshold falling edge position". Finally, according to the signal-to-noise ratio estimation result, the Select one of "Falling Edge Position" and "Wide Time Window Falling Edge Position".

Further, the specific steps of the pulse width detection algorithm of radar and communication signals are:

(1) Parameter initialization settings:

(1-1) Set a wide time window and divide the wide time window into K narrow time windows, the length of the wide time window is N sampling points, the sampling interval is Ts, N is greater than or equal to K, and the narrow time window The length of the time window is M sampling points, and M is less than N;

(1-2) Set a signal-to-noise ratio value SNR0;

(1-3) Set false alarm probability: set false alarm probability Pf0 for wide time window; set false alarm probability Pf1 and Pf2 for narrow time window, Pf0<Pf1<Pf2;

(1-4) Calculate each detection threshold according to each false alarm probability: calculate the detection threshold T0 of the wide time window according to Pf0, calculate the detection threshold T1 and T2 of the narrow time window according to Pf1 and Pf2, and have T0>T1>T2 , that is, T1 is the high threshold of the narrow time window, and T2 is the low threshold of the narrow time window;

(1-5) Set the "limit of sampling points for estimating the signal-to-noise ratio" N0;

(2) Clear the internal input signal sampling values and sampling value buffer queues, input signal power and power buffer queues of all wide time windows and narrow time windows;

Establish the flag X1 of "the rising edge of the next pulse is detected in the falling edge detection", the flag X2 of the "falling edge of the pulse detected in the rising edge detection", the flag X3 of "the pulse signal has been detected in a wide time window" and the flag "not detected To the pulse signal" flag X4 and set all to 0; establish "the number of sampling points with wide time window power below the threshold" N1 and "number of sampling points with wide time window over the threshold" N2 and set them to 0;

(3) Run the rising edge detection module:

(3-1) Set up the "low threshold rising edge detected" flag Y1 of the narrow time window and the "high threshold rising edge detected" flag Y2 of the narrow time window and set them all to 0;

(3-2) Check whether the flag X1 is 1: if it is 1, it means that the rising edge of the current pulse is detected in the falling edge detection module of the previous pulse, then set the flag X1 to 0, and then enter step (3-4), Detect the rising edge position according to the data in the buffer area, the data in the buffer area includes the average power of the wide time window and each narrow time window, and also includes the data in the buffer queue; if it is not 1, then enter the step (3- 3);

(3-3) Calculate the average power of the wide time window and the average power of K narrow time windows, and process the cache queue corresponding to the wide time window and each narrow time window: the current sampling point enters the cache queue and the cache queue The last sampling point exits the cache queue;

(3-4) Determine whether the average power of each narrow time window is greater than or equal to T1, and obtain K high-threshold judgment results; and separately judge whether the average power of each narrow time window is greater than or equal to T2, and obtain K low-threshold judgment results ;

(3-5) Establish the "rising edge low-threshold fusion decision" flag Z1, and then perform decision fusion on the low-threshold judgment results of K narrow time windows; if the fusion result is that a pulse signal is detected, set Z1 to 1, otherwise Set Z1 to 0;

(3-6) Establish the "rising edge high threshold fusion decision" flag Z2, and then perform decision fusion on the high threshold judgment results of K narrow time windows; if the fusion result is that a pulse signal is detected, set Z2 to 1, otherwise Set Z2 to 0;

(3-7) Check whether the flag X3 is 1: if the flag X3 is 1, add 1 to N2, and then enter step (3-8); if the flag X3 is not 1, then judge whether the average power of the wide time window is greater than Equal to T0: If the average power of the wide time window is greater than or equal to T0, set the flag X3 to 1, set N2 to 1, and record the "rising edge position of the wide time window", and then enter step (3-8), if the wide time window If the average power is less than T0, enter step (3-3) to process the next sampling point;

(3-8) Perform second-level decision fusion on the wide time window detection result and the rising edge low threshold fusion decision, and check that the pulse signal is detected for the first time during the rising edge low threshold fusion decision process after the pulse signal is detected in the wide time window moment: if the flag Y1 is 0, check whether the flag Z1 is 1: if the flag Z1 is 1, set the flag Y1 to 1 and record the "low threshold rising edge position" and then enter step (3-9), if the flag Z1 If it is not 1, go to step (3-9); if the flag Y1 is not 0, go to step (3-9);

(3-9) Perform second-level decision fusion on the wide time window detection result and the rising edge high threshold fusion decision, and check that the pulse signal is detected for the first time during the rising edge high threshold fusion decision process after the pulse signal is detected in the wide time window moment: if the flag Y2 is 0, check whether the flag Z2 is 1: if the flag Z2 is 1, set the flag Y2 to 1 and record the "high threshold rising edge position" and then enter step (3-10), if the flag Z2 If it is not 1, go to step (3-10); if the flag Y2 is not 0, go to step (3-10);

(3-10) After the pulse signal is detected in the wide time window, check whether the falling edge of the current pulse appears: if the average power of the current wide time window is less than T0, it indicates that the pulse signal is detected after the pulse signal is detected in the wide time window during the rising edge detection process It disappears again, that is, the falling edge of the current pulse arrives, then record the "wide time window falling edge position" and set the flags X2 and X4 to 1, and set the flags X3 and N1 to 0, and then execute the "adaptive rising edge position Confirm" module and give the position of the rising edge of the current pulse, then complete the rising edge detection and return; if the average power of the current wide time window is not less than T0, estimate the signal-to-noise ratio SNR, and then enter step (3-11);

(3-11) Check whether the rising edge detection is completed: after the pulse signal is detected in a wide time window, if N2 is greater than N0, it indicates that the rising edge detection and signal-to-noise ratio estimation have been completed, and the flags X1, X2, X3 and X4 are all Set to 0, then execute the "adaptive rising edge position determination" module and give the rising edge position, and finally the rising edge detection is completed and returned; if N2 is not greater than N0, then jump to step (3-3) for the next sampling point process;

The steps of the "adaptive rising edge position determination" module are:

(A-1) If SNR>=SNR0 and the flag Y2 is 1, set the rising edge position as "high threshold rising edge position" and return, otherwise enter step (A-2);

(A-2) If SNR<SNR0 and the flag Y1 is 1, set the rising edge position to "low threshold rising edge position" and return, otherwise enter step (A-3);

(A-3) Set the rising edge position as "Wide Time Window Rising Edge Position", and then return;

(4) Run the falling edge detection module:

(4-1) Set up the "low threshold falling edge detected" flag Y3 of the narrow time window and the "high threshold falling edge detected" flag Y4 of the narrow time window and set them all to 0;

(4-2) Check whether the flag X2 is 1: if it is 1, it means that the pulse signal disappears after appearing in the rising edge detection module, then set the flag X2 to 0, and then enter step (4-4), according to the The position of the falling edge of the data detection, the data in the buffer area includes the average power of the wide time window and each narrow time window, and also includes the data in the buffer queue; if it is not 1, enter step (4-3);

(4-3) Calculate the average power of the wide time window and the respective average power of K narrow time windows, and process the buffer queue corresponding to the wide time window and each narrow time window: the current sampling point enters the buffer queue and buffer queue The last sampling point exits the cache queue;

(4-4) Determine whether the average power of each narrow time window is lower than T1, and obtain K high-threshold judgment results; and respectively judge whether the average power of each narrow time window is lower than T2, and obtain K low-threshold judgment results ;

(4-5) Establish the "falling edge low threshold fusion decision" flag Z3, and then perform decision fusion on the low threshold judgment results of K narrow time windows; if the fusion result is that no pulse signal is detected, set Z3 to 1, Otherwise set Z3 to 0;

(4-6) Establish the "falling edge high threshold fusion decision" flag Z4, and then perform decision fusion on the high threshold judgment results of K narrow time windows; if the fusion result is that no pulse signal is detected, set Z4 to 1, Otherwise set Z4 to 0;

(4-7) Check whether the flag X4 is 0: if the flag X4 is 0, go to step (4-8); if the flag X4 is not 0, add 1 to N1 and go to step (4-9);

(4-8) Determine whether the average power of the wide time window is lower than T0: If the average power of the wide time window is lower than T0, set both flags X4 and N1 to 1, and record the "falling edge position of the wide time window", and then Enter (4-9); if the average power of the wide time window is greater than or equal to T0, estimate the signal-to-noise ratio SNR, and then enter step (4-3) to process the next sampling point;

(4-9) The detection result of the wide time window and the low-threshold fusion decision of the falling edge are used for the second-level decision fusion, and the pulse signal is detected for the first time in the process of fusion decision-making process of the low-threshold fusion of the falling edge after the wide time window detects the disappearance of the pulse signal The moment of disappearance: if the flag Y3 is 0, check whether the flag Z3 is 1: if the flag Z3 is 1, set the flag Y3 to 1 and record the "low threshold falling edge position" and then enter step (4-10), if the flag If Z3 is not 1, enter step (4-10); if the flag Y3 is not 0, enter step (4-10);

(4-10) Perform second-level decision fusion between the detection results of the wide time window and the high-threshold fusion decision on the falling edge, and check that the pulse is detected for the first time during the high-threshold fusion decision-making process of the falling edge after the pulse signal disappears in the wide time window The moment when the signal disappears: if the flag Y4 is 0, check whether the flag Z4 is 1: if the flag Z4 is 1, set the flag Y4 to 1 and record the "high threshold falling edge position" and then enter step (4-11), if If the flag Z4 is not 1, go to step (4-11); if the flag Y4 is not 0, go to step (4-11);

(4-11) Check whether the rising edge of the next pulse appears: if the average power of the current wide time window is greater than or equal to T0, it indicates that a new pulse signal is detected during the falling edge detection process, that is, the rising edge of the next pulse arrives, Then record the "rising edge position of wide time window" and set the flags X1 and X3 to 1, and set the flags X4 and N2 to 0, then execute the "adaptive falling edge position determination" module and give the current pulse falling edge position, and then The falling edge detection is completed and returns; if the average power of the current wide time window is less than T0, enter step (4-12);

(4-12) Check whether the falling edge detection is completed: if N1 is greater than N, set the flags X1, X2, X3 and flag X4 to 0, then execute the "adaptive falling edge position determination" module and give the falling edge position, and then The falling edge detection is completed and returns; if N1 is less than or equal to N, then jump to step (4-3) to process the next sampling point;

(5) Calculate the pulse width according to the position of the rising edge and the position of the falling edge, and then enter step (3) to start detecting the rising edge again;

The steps of the "adaptive falling edge position determination" module are:

(B-1) If SNR>=SNR0 and the flag Y4 is 1, then the falling edge position is "high threshold falling edge position", and then return, otherwise enter step (B-2);

(B-2) If SNR<SNR0 and flag Y3 is 1, then the falling edge position is "low threshold falling edge position", and then return, otherwise enter step (B-3);

(B-3) Set the falling edge time as the "wide time window falling edge position", and then return.

Compared with the prior art, the present invention has the following advantages: (1) The combination of wide window and narrow window is adopted, the number of sampling points contained in the two is different, and different false alarm probabilities, namely threshold values, are set. The window sets a very low false alarm probability, while the narrow window sets a high false alarm probability, so that both can pass the threshold at the same time, balancing the relationship between detection probability and positioning rising and falling edges, taking advantage of the advantages of wide windows The detection ability ensures extremely low pulse loss probability and extremely low false alarm probability. At the same time, it uses the high resolution capability of narrow window to improve the detection accuracy of rising and falling edges; (2) adopts two-level decision fusion mechanism, multiple The fusion result after a narrow window decision fusion is fused with the judgment result of the wide window, which further improves the detection probability and detection accuracy; (3) The threshold value of the narrow window can be adjusted and switched adaptively according to the signal-to-noise ratio estimation result, and the detection High precision, strong flexibility, and wide applicability; (4) The main operation is energy (or power) calculation and decision fusion, and the algorithm operation complexity is low and has nothing to do with the specific waveform shape; (5) Fully consider the rising edge detection process The falling edge is detected again and the rising edge is detected during the falling edge detection process, which improves the detection accuracy and broadens the scope of application.

Description of drawings

Fig. 1 is the overall schematic flow chart of the present invention.

Fig. 2 is the first part of the schematic flow chart of rising edge detection.

Fig. 3 is the second part of the schematic flow chart of rising edge detection, wherein node 1.1 and node 1.2 correspond to node 1.1 and node 1.2 in Fig. 2 respectively.

Fig. 4 is a flow chart of the "adaptive rising edge position determination" module.

FIG. 5 is the first part of the flow diagram of falling edge detection.

FIG. 6 is the second part of the flow diagram of falling edge detection, wherein the node 2.1 and the node 2.2 correspond to the node 2.1 and the node 2.2 in FIG. 5 respectively.

Fig. 7 is a flow chart of the "adaptive falling edge position determination" module.

detailed description

The technical scheme of the present invention is described in detail below in conjunction with accompanying drawing:

A pulse width detection algorithm for radar and communication signals, the main steps are: first detect the position of the rising edge, then detect the position of the falling edge, then obtain the pulse width through the position of the rising edge and the position of the falling edge, and then enter the detection of the rising edge again , so that the detection of the rising edge and the detection of the falling edge are cyclically carried out, and the pulse signal is continuously captured and the pulse width is calculated, which is characterized in that:

Before detection, first establish a wide time window, then divide the wide time window into several narrow time windows, and set a wide time window threshold for the wide time window, and set a high threshold and a low threshold for the narrow time window;

When the rising edge is detected, the judgment of the over-high threshold and the under-threshold of the narrow time window are first carried out (note: the over-threshold mentioned in the present invention refers to "the average power in the time window is greater than or equal to the threshold"), and then the high-threshold decision fusion is carried out It is fused with the low-threshold decision, and then the threshold crossing judgment of the wide time window is carried out to obtain the "rising edge position of the wide time window". Threshold rising edge position", the second-level decision-making fusion of the wide time window detection result and the rising edge high threshold fusion decision is obtained to obtain the "high threshold rising edge position", and the signal-to-noise ratio is estimated for the sampling points that pass the threshold in the wide time window, Finally, select one of the "rising edge position" from "low threshold rising edge position", "high threshold rising edge position" and "wide time window rising edge position" according to the result of signal-to-noise ratio estimation;

When detecting the falling edge, first judge the over-high threshold and the low-threshold judgment of the narrow time window, then perform the high-threshold decision fusion and low-threshold decision fusion, and then carry out the wide time window cross-threshold judgment to complete the wide time window power greater than or equal to The signal-to-noise ratio of the sampling point of the threshold is estimated to obtain the "wide time window falling edge position", and then the wide time window detection result is combined with the falling edge low threshold fusion decision for the second-level decision fusion to obtain the "low threshold falling edge position", The second-level decision fusion is performed on the wide time window detection result and the falling edge high threshold fusion decision to obtain the "high threshold falling edge position". Finally, according to the signal-to-noise ratio estimation result, the Select one of "Falling Edge Position" and "Wide Time Window Falling Edge Position".

The specific steps of this pulse width detection algorithm are:

As shown in Figure 1, (1) Parameter initialization settings:

(1-1) Set a wide time window and divide the wide time window into K narrow time windows, the length of the wide time window is N sampling points, the sampling interval is Ts, N is greater than or equal to K, and the narrow time window The length of the time window is M sampling points, and M is less than N;

(1-2) Set a signal-to-noise ratio value SNR0;

(1-3) Set false alarm probability: set false alarm probability Pf0 for wide time window; set false alarm probability Pf1 and Pf2 for narrow time window, Pf0<Pf1<Pf2; Pf0 is one or several times lower than the other two order of magnitude, such as: Pf=0.000001, Pf1=0.0001, Pf2=0.01;

(1-4) Calculate each detection threshold according to each false alarm probability: calculate the detection threshold T0 of the wide time window according to Pf0, calculate the detection threshold T1 and T2 of the narrow time window according to Pf1 and Pf2, and have T0>T1>T2 , that is, T1 is the high threshold of the narrow time window, and T2 is the low threshold of the narrow time window;

(1-5) Set the "limit of sampling points for estimating the signal-to-noise ratio" N0;

(2) Clear the internal input signal sampling values and sampling value buffer queues, input signal power and power buffer queues of all wide time windows and narrow time windows;

Establish the flag X1 of "the rising edge of the next pulse is detected in the falling edge detection", the flag X2 of the "falling edge of the pulse detected in the rising edge detection", the flag X3 of "the pulse signal has been detected in a wide time window" and the flag "not detected To the pulse signal" flag X4 and set all to 0; establish "the number of sampling points with wide time window power below the threshold" N1 and "number of sampling points with wide time window over the threshold" N2 and set them to 0;

(3) Run the rising edge detection module:

(3-1) As shown in Figure 2, set up the "low threshold rising edge detected" flag Y1 of the narrow time window and the "high threshold rising edge detected" flag Y2 of the narrow time window and set them all to 0;

(3-2) Check whether the flag X1 is 1: if it is 1, it means that the rising edge of the current pulse is detected in the falling edge detection module of the previous pulse, then set the flag X1 to 0, and then enter step (3-4), Detect the rising edge position according to the data in the buffer area, the data in the buffer area includes the average power of the wide time window and each narrow time window, and also includes the data in the buffer queue; if it is not 1, then enter the step (3- 3);

(3-3) Calculate the average power of the wide time window and the average power of K narrow time windows, and process the cache queue corresponding to the wide time window and each narrow time window: the current sampling point enters the cache queue and the cache queue The last sampling point exits the cache queue;

(3-4) Determine whether the average power of each narrow time window is greater than or equal to T1, and obtain K high-threshold judgment results; and separately judge whether the average power of each narrow time window is greater than or equal to T2, and obtain K low-threshold judgment results ;

(3-5) Establish the "rising edge low-threshold fusion decision" flag Z1, and then perform decision fusion on the low-threshold judgment results of K narrow time windows; if the fusion result is that a pulse signal is detected, set Z1 to 1, otherwise Set Z1 to 0;

(3-6) Establish the "rising edge high threshold fusion decision" flag Z2, and then perform decision fusion on the high threshold judgment results of K narrow time windows; if the fusion result is that a pulse signal is detected, set Z2 to 1, otherwise Set Z2 to 0;

(3-7) Check whether the flag X3 is 1: if the flag X3 is 1, add 1 to N2, and then enter step (3-8); if the flag X3 is not 1, then judge whether the average power of the wide time window is greater than Equal to T0: If the average power of the wide time window is greater than or equal to T0, set the flag X3 to 1, set N2 to 1, and record the "rising edge position of the wide time window", and then enter step (3-8), if the wide time window If the average power is less than T0, enter step (3-3) to process the next sampling point;

(3-8) Perform second-level decision fusion on the wide time window detection result and the rising edge low threshold fusion decision, and check that the pulse signal is detected for the first time during the rising edge low threshold fusion decision process after the pulse signal is detected in the wide time window moment: if the flag Y1 is 0, check whether the flag Z1 is 1: if the flag Z1 is 1, set the flag Y1 to 1 and record the "low threshold rising edge position" and then enter step (3-9), if the flag Z1 If it is not 1, go to step (3-9); if the flag Y1 is not 0, go to step (3-9);

(3-9) Perform second-level decision fusion on the wide time window detection result and the rising edge high threshold fusion decision, and check that the pulse signal is detected for the first time during the rising edge high threshold fusion decision process after the pulse signal is detected in the wide time window moment: if the flag Y2 is 0, check whether the flag Z2 is 1: if the flag Z2 is 1, set the flag Y2 to 1 and record the "high threshold rising edge position" and then enter step (3-10), if the flag Z2 If it is not 1, go to step (3-10); if the flag Y2 is not 0, go to step (3-10);

(3-10) After the pulse signal is detected in the wide time window, check whether the falling edge of the current pulse appears: if the average power of the current wide time window is less than T0, it indicates that the pulse signal is detected after the pulse signal is detected in the wide time window during the rising edge detection process It disappears again, that is, the falling edge of the current pulse arrives, then record the "wide time window falling edge position" and set the flags X2 and X4 to 1, and set the flags X3 and N1 to 0, and then execute the "adaptive rising edge position Confirm" module and give the position of the rising edge of the current pulse, then complete the rising edge detection and return; if the average power of the current wide time window is not less than T0, estimate the signal-to-noise ratio SNR, and then enter step (3-11);

(3-11) Check whether the rising edge detection is completed: after the pulse signal is detected in a wide time window, if N2 is greater than N0, it indicates that the rising edge detection and signal-to-noise ratio estimation have been completed, and the flags X1, X2, X3 and X4 are all Set to 0, then execute the "adaptive rising edge position determination" module and give the rising edge position, and finally the rising edge detection is completed and returned; if N2 is not greater than N0, then jump to step (3-3) for the next sampling point process;

As shown in Figure 4, the steps of the "adaptive rising edge position determination" module are:

(A-1) If SNR>=SNR0 and the flag Y2 is 1, set the rising edge position as "high threshold rising edge position" and return, otherwise enter step (A-2);

(A-2) If SNR<SNR0 and the flag Y1 is 1, set the rising edge position to "low threshold rising edge position" and return, otherwise enter step (A-3);

(A-3) Set the rising edge position as "Wide Time Window Rising Edge Position", and then return;

(4) As shown in Figure 1, run the falling edge detection module:

(4-1) As shown in Figure 5, set up the "low threshold falling edge detected" flag Y3 of the narrow time window and the "high threshold falling edge detected" flag Y4 of the narrow time window and set them all to 0;

(4-2) Check whether the flag X2 is 1: if it is 1, it means that the pulse signal disappears after appearing in the rising edge detection module, then set the flag X2 to 0, and then enter step (4-4), according to the The position of the falling edge of the data detection, the data in the buffer area includes the average power of the wide time window and each narrow time window, and also includes the data in the buffer queue; if it is not 1, enter step (4-3);

(4-3) Calculate the average power of the wide time window and the respective average power of K narrow time windows, and process the buffer queue corresponding to the wide time window and each narrow time window: the current sampling point enters the buffer queue and buffer queue The last sampling point exits the cache queue;

(4-4) Determine whether the average power of each narrow time window is lower than T1, and obtain K high-threshold judgment results; and respectively judge whether the average power of each narrow time window is lower than T2, and obtain K low-threshold judgment results ;

(4-5) Establish the "falling edge low threshold fusion decision" flag Z3, and then perform decision fusion on the low threshold judgment results of K narrow time windows; if the fusion result is that no pulse signal is detected, set Z3 to 1, Otherwise set Z3 to 0;

(4-6) Establish the "falling edge high threshold fusion decision" flag Z4, and then perform decision fusion on the high threshold judgment results of K narrow time windows; if the fusion result is that no pulse signal is detected, set Z4 to 1, Otherwise set Z4 to 0;

(4-7) Check whether the flag X4 is 0: if the flag X4 is 0, go to step (4-8); if the flag X4 is not 0, add 1 to N1 and go to step (4-9);

(4-8) Determine whether the average power of the wide time window is lower than T0: If the average power of the wide time window is lower than T0, set both flags X4 and N1 to 1, and record the "falling edge position of the wide time window", and then Enter (4-9); if the average power of the wide time window is greater than or equal to T0, estimate the signal-to-noise ratio SNR, and then enter step (4-3) to process the next sampling point;

(4-9) As shown in Figure 6, the detection results of the wide time window and the low-threshold fusion decision of the falling edge are used for the second-level decision fusion, and the first time in the process of the low-threshold fusion decision-making process of the falling edge is checked after the pulse signal disappears in the wide time window. When the pulse signal disappears is detected: if the flag Y3 is 0, check whether the flag Z3 is 1: if the flag Z3 is 1, set the flag Y3 to 1 and record the "low threshold falling edge position" and then enter the step (4-10 ), if the flag Z3 is not 1, then enter step (4-10); if the flag Y3 is not 0, then enter step (4-10);

(4-10) Perform second-level decision fusion between the detection results of the wide time window and the high-threshold fusion decision on the falling edge, and check that the pulse is detected for the first time during the high-threshold fusion decision-making process of the falling edge after the pulse signal disappears in the wide time window The moment when the signal disappears: if the flag Y4 is 0, check whether the flag Z4 is 1: if the flag Z4 is 1, set the flag Y4 to 1 and record the "high threshold falling edge position" and then enter step (4-11), if If the flag Z4 is not 1, go to step (4-11); if the flag Y4 is not 0, go to step (4-11);

(4-11) Check whether the rising edge of the next pulse appears: if the average power of the current wide time window is greater than or equal to T0, it indicates that a new pulse signal is detected during the falling edge detection process, that is, the rising edge of the next pulse arrives, Then record the "rising edge position of wide time window" and set the flags X1 and X3 to 1, and set the flags X4 and N2 to 0, then execute the "adaptive falling edge position determination" module and give the current pulse falling edge position, and then The falling edge detection is completed and returns; if the average power of the current wide time window is less than T0, enter step (4-12);

(4-12) Check whether the falling edge detection is completed: if N1 is greater than N, set the flags X1, X2, X3 and flag X4 to 0, then execute the "adaptive falling edge position determination" module and give the falling edge position, and then The falling edge detection is completed and returns; if N1 is less than or equal to N, then jump to step (4-3) to process the next sampling point;

(5) As shown in Figure 1, calculate the pulse width according to the position of the rising edge and the position of the falling edge, and then enter step (3) to start detecting the rising edge again.

As shown in Figure 7, the steps of the "adaptive falling edge position determination" module are:

(B-1) If SNR>=SNR0 and the flag Y4 is 1, then the falling edge position is "high threshold falling edge position", and then return, otherwise enter step (B-2);

(B-2) If SNR<SNR0 and flag Y3 is 1, then the falling edge position is "low threshold falling edge position", and then return, otherwise enter step (B-3);

(B-3) Set the falling edge time as the "wide time window falling edge position", and then return.

Claims (2)

1. A pulse width detection algorithm for radar and communication signals, the steps are: first detect the rising position, then detect the falling edge position, then obtain the pulse width through the rising edge position and the falling edge position, and then enter the rising edge detection again , so that the detection of the rising edge and the detection of the falling edge are cyclically carried out, and the pulse signal is continuously captured and the pulse width is calculated, which is characterized in that: Before detection, first establish a wide time window, then divide the wide time window into several narrow time windows, and set a wide time window threshold for the wide time window, and set a high threshold and a low threshold for the narrow time window; When detecting the rising edge, firstly judge the over-high threshold and the low-threshold judgment of the narrow time window, then carry out the high-threshold decision fusion and the low-threshold decision fusion, and then carry out the wide time window cross-threshold judgment and get the "wide time window rising edge position", and then perform second-level decision fusion on the wide time window detection result and the rising edge low threshold fusion decision to obtain the "low threshold rising edge position", and perform the second level decision on the wide time window detection result and the rising edge high threshold fusion decision Decision fusion and obtain the "high threshold rising edge position", and estimate the signal-to-noise ratio of the sampling points that pass the threshold in a wide time window. Finally, according to the result of the signal-to-noise ratio estimation, the " and "Wide Time Window Rising Edge Position" select one as "Rising Edge Position"; When detecting the falling edge, first judge the over-high threshold and the low-threshold judgment of the narrow time window, then perform the high-threshold decision fusion and low-threshold decision fusion, and then carry out the wide time window cross-threshold judgment to complete the wide time window power greater than or equal to The signal-to-noise ratio of the sampling point of the threshold is estimated to obtain the "wide time window falling edge position", and then the wide time window detection result is combined with the falling edge low threshold fusion decision for the second-level decision fusion to obtain the "low threshold falling edge position", The second-level decision fusion is performed on the wide time window detection result and the falling edge high threshold fusion decision to obtain the "high threshold falling edge position". Finally, according to the signal-to-noise ratio estimation result, the Select one of "Falling Edge Position" and "Wide Time Window Falling Edge Position". 2. The pulse width detection algorithm of radar and communication signal as claimed in claim 1, is characterized in that concrete steps are: (1) Parameter initialization settings: (1-1) Set a wide time window and divide the wide time window into K narrow time windows, the length of the wide time window is N sampling points, the sampling interval is Ts, N is greater than or equal to K, and the narrow time window The length of the time window is M sampling points, and M is less than N; (1-2) Set a signal-to-noise ratio value SNR0; (1-3) Set false alarm probability: set false alarm probability Pf0 for wide time window; set false alarm probability Pf1 and Pf2 for narrow time window, Pf0<Pf1<Pf2; (1-4) Calculate each detection threshold according to each false alarm probability: calculate the detection threshold T0 of the wide time window according to Pf0, calculate the detection threshold T1 and T2 of the narrow time window according to Pf1 and Pf2, and have T0>T1>T2 , that is, T1 is the high threshold of the narrow time window, and T2 is the low threshold of the narrow time window; (1-5) Set the "limit of sampling points for estimating the signal-to-noise ratio" N0; (2) Clear the internal input signal sampling values and sampling value buffer queues, input signal power and power buffer queues of all wide time windows and narrow time windows; Establish the flag X1 of "the rising edge of the next pulse is detected in the falling edge detection", the flag X2 of the "falling edge of the pulse detected in the rising edge detection", the flag X3 of "the pulse signal has been detected in a wide time window" and the flag "not detected To the pulse signal" flag X4 and set all to 0; establish "the number of sampling points with wide time window power below the threshold" N1 and "number of sampling points with wide time window over the threshold" N2 and set them to 0; (3) Run the rising edge detection module: (3-1) Set up the "low threshold rising edge detected" flag Y1 of the narrow time window and the "high threshold rising edge detected" flag Y2 of the narrow time window and set them all to 0; (3-2) Check whether the flag X1 is 1: if it is 1, it means that the rising edge of the current pulse is detected in the falling edge detection module of the previous pulse, then set the flag X1 to 0, and then enter step (3-4), Detect the rising edge position according to the data in the buffer area, the data in the buffer area includes the average power of the wide time window and each narrow time window, and also includes the data in the buffer queue; if it is not 1, then enter the step (3- 3); (3-3) Calculate the average power of the wide time window and the average power of K narrow time windows, and process the cache queue corresponding to the wide time window and each narrow time window: the current sampling point enters the cache queue and the cache queue The last sampling point exits the cache queue; (3-4) Determine whether the average power of each narrow time window is greater than or equal to T1, and obtain K high-threshold judgment results; and separately judge whether the average power of each narrow time window is greater than or equal to T2, and obtain K low-threshold judgment results ; (3-5) Establish the "rising edge low-threshold fusion decision" flag Z1, and then perform decision fusion on the low-threshold judgment results of K narrow time windows; if the fusion result is that a pulse signal is detected, set Z1 to 1, otherwise Set Z1 to 0; (3-6) Establish the "rising edge high threshold fusion decision" flag Z2, and then perform decision fusion on the high threshold judgment results of K narrow time windows; if the fusion result is that a pulse signal is detected, set Z2 to 1, otherwise Set Z2 to 0; (3-7) Check whether the flag X3 is 1: if the flag X3 is 1, add 1 to N2, and then enter step (3-8); if the flag X3 is not 1, then judge whether the average power of the wide time window is greater than Equal to T0: If the average power of the wide time window is greater than or equal to T0, set the flag X3 to 1, set N2 to 1, and record the "rising edge position of the wide time window", and then enter step (3-8), if the wide time window If the average power is less than T0, enter step (3-3) to process the next sampling point; (3-8) Perform second-level decision fusion on the wide time window detection result and the rising edge low threshold fusion decision, and check that the pulse signal is detected for the first time during the rising edge low threshold fusion decision process after the pulse signal is detected in the wide time window moment: if the flag Y1 is 0, check whether the flag Z1 is 1: if the flag Z1 is 1, set the flag Y1 to 1 and record the "low threshold rising edge position" and then enter step (3-9), if the flag Z1 If it is not 1, go to step (3-9); if the flag Y1 is not 0, go to step (3-9); (3-9) Perform second-level decision fusion on the wide time window detection result and the rising edge high threshold fusion decision, and check that the pulse signal is detected for the first time during the rising edge high threshold fusion decision process after the pulse signal is detected in the wide time window moment: if the flag Y2 is 0, check whether the flag Z2 is 1: if the flag Z2 is 1, set the flag Y2 to 1 and record the "high threshold rising edge position" and then enter step (3-10), if the flag Z2 If it is not 1, go to step (3-10); if the flag Y2 is not 0, go to step (3-10); (3-10) After the pulse signal is detected in the wide time window, check whether the falling edge of the current pulse appears: if the average power of the current wide time window is less than T0, it indicates that the pulse signal is detected after the pulse signal is detected in the wide time window during the rising edge detection process It disappears again, that is, the falling edge of the current pulse arrives, then record the "wide time window falling edge position" and set the flags X2 and X4 to 1, and set the flags X3 and N1 to 0, and then execute the "adaptive rising edge position Confirm" module and give the position of the rising edge of the current pulse, then complete the rising edge detection and return; if the average power of the current wide time window is not less than T0, estimate the signal-to-noise ratio SNR, and then enter step (3-11); (3-11) Check whether the rising edge detection is completed: after the pulse signal is detected in a wide time window, if N2 is greater than N0, it indicates that the rising edge detection and signal-to-noise ratio estimation have been completed, and the flags X1, X2, X3 and X4 are all Set to 0, then execute the "adaptive rising edge position determination" module and give the rising edge position, and finally the rising edge detection is completed and returned; if N2 is not greater than N0, then jump to step (3-3) for the next sampling point process; The steps of the "adaptive rising edge position determination" module are: (A-1) If SNR>=SNR0 and the flag Y2 is 1, set the rising edge position as "high threshold rising edge position" and return, otherwise enter step (A-2); (A-2) If SNR<SNR0 and the flag Y1 is 1, set the rising edge position to "low threshold rising edge position" and return, otherwise enter step (A-3); (A-3) Set the rising edge position as "Wide Time Window Rising Edge Position", and then return; (4) Run the falling edge detection module: (4-1) Set up the "low threshold falling edge detected" flag Y3 of the narrow time window and the "high threshold falling edge detected" flag Y4 of the narrow time window and set them all to 0; (4-2) Check whether the flag X2 is 1: if it is 1, it means that the pulse signal disappears after appearing in the rising edge detection module, then set the flag X2 to 0, and then enter step (4-4), according to the The position of the falling edge of the data detection, the data in the buffer area includes the average power of the wide time window and each narrow time window, and also includes the data in the buffer queue; if it is not 1, enter step (4-3); (4-3) Calculate the average power of the wide time window and the respective average power of K narrow time windows, and process the buffer queue corresponding to the wide time window and each narrow time window: the current sampling point enters the buffer queue and buffer queue The last sampling point exits the cache queue; (4-4) Determine whether the average power of each narrow time window is lower than T1, and obtain K high-threshold judgment results; and respectively judge whether the average power of each narrow time window is lower than T2, and obtain K low-threshold judgment results ; (4-5) Establish the "falling edge low threshold fusion decision" flag Z3, and then perform decision fusion on the low threshold judgment results of K narrow time windows; if the fusion result is that no pulse signal is detected, set Z3 to 1, Otherwise set Z3 to 0; (4-6) Establish the "falling edge high threshold fusion decision" flag Z4, and then perform decision fusion on the high threshold judgment results of K narrow time windows; if the fusion result is that no pulse signal is detected, set Z4 to 1, Otherwise set Z4 to 0; (4-7) Check whether the flag X4 is 0: if the flag X4 is 0, go to step (4-8); if the flag X4 is not 0, add 1 to N1 and go to step (4-9); (4-8) Determine whether the average power of the wide time window is lower than T0: If the average power of the wide time window is lower than T0, set both flags X4 and N1 to 1, and record the "falling edge position of the wide time window", and then Enter (4-9); if the average power of the wide time window is greater than or equal to T0, estimate the signal-to-noise ratio SNR, and then enter step (4-3) to process the next sampling point; (4-9) The detection result of the wide time window and the low-threshold fusion decision of the falling edge are used for the second-level decision fusion, and the pulse signal is detected for the first time in the process of fusion decision-making process of the low-threshold fusion of the falling edge after the wide time window detects the disappearance of the pulse signal The moment of disappearance: if the flag Y3 is 0, check whether the flag Z3 is 1: if the flag Z3 is 1, set the flag Y3 to 1 and record the "low threshold falling edge position" and then enter step (4-10), if the flag If Z3 is not 1, enter step (4-10); if the flag Y3 is not 0, enter step (4-10); (4-10) Perform second-level decision fusion between the detection results of the wide time window and the high-threshold fusion decision on the falling edge, and check that the pulse is detected for the first time during the high-threshold fusion decision-making process of the falling edge after the pulse signal disappears in the wide time window The moment when the signal disappears: if the flag Y4 is 0, check whether the flag Z4 is 1: if the flag Z4 is 1, set the flag Y4 to 1 and record the "high threshold falling edge position" and then enter step (4-11), if If the flag Z4 is not 1, go to step (4-11); if the flag Y4 is not 0, go to step (4-11); (4-11) Check whether the rising edge of the next pulse appears: if the average power of the current wide time window is greater than or equal to T0, it indicates that a new pulse signal is detected during the falling edge detection process, that is, the rising edge of the next pulse arrives, Then record the "rising edge position of wide time window" and set the flags X1 and X3 to 1, and set the flags X4 and N2 to 0, then execute the "adaptive falling edge position determination" module and give the current pulse falling edge position, and then The falling edge detection is completed and returns; if the average power of the current wide time window is less than T0, enter step (4-12); (4-12) Check whether the falling edge detection is completed: if N1 is greater than N, set the flags X1, X2, X3 and flag X4 to 0, then execute the "adaptive falling edge position determination" module and give the falling edge position, and then The falling edge detection is completed and returns; if N1 is less than or equal to N, then jump to step (4-3) to process the next sampling point; (5) Calculate the pulse width according to the position of the rising edge and the position of the falling edge, and then enter step (3) to start detecting the rising edge again; The steps of the "adaptive falling edge position determination" module are: (B-1) If SNR>=SNR0 and the flag Y4 is 1, then the falling edge position is "high threshold falling edge position", and then return, otherwise enter step (B-2); (B-2) If SNR<SNR0 and flag Y3 is 1, then the falling edge position is "low threshold falling edge position", and then return, otherwise enter step (B-3); (B-3) Set the falling edge time as the "wide time window falling edge position", and then return.